Conservation of structure, signaling and pharmacology between two serotonin receptor subtypes from decapod crustaceans, Panulirus interruptus and Procambarus clarkii

Serotonin (5-HT) is an important neurotransmitter and neurohormone in most animal species. Diverse processes, including social behaviors and a variety of systemic physiological functions are regulated by this biogenic amine. Animals have evolved a suite of 5-HT receptor types; thus, a ingle neurotransmitter can have multiple and complex effects on different tissues. To date, seven classes of serotonin receptors have been identified in mammals. These receptor classes are categorized with respect to their signal transduction mechanisms and pharmacological properties (Gerhardt and van Heerikhuizen, 1997; Hoyer et al., 2002; Kroeze et al., 2002).

In invertebrate nervous systems 5-HT can modulate motor pattern generation(Hooper and DiCaprio, 2004),escape and social status (Edwards et al.,1999), aggression (Kravitz,2000) and learning (Barbas et al., 2003; Bicker,1999; Kandel and Schwartz,1982). Serotonin has multiple and complex roles in crustacean models. 5-HT application to the stomatogastric nervous system of decapod crustaceans elicits distinct responses from individual identified neurons(Flamm and Harris-Warrick,1986), and these responses can differ for the same identified neuron in different species (Katz and Tazaki, 1992). In the crayfish, the social status of an individual determines the directionality of 5-HT modulation of the lateral giant escape circuit response to sensory stimulation(Yeh et al., 1997; Yeh et al., 1996). The rate and concentration of 5-HT application to the circuit also affect this response(Teshiba et al., 2001). These studies suggest that crustaceans express several different 5-HT receptor types and that the expression or signaling of these receptors might change in response to social or environmental stimuli.

Crustacean hormonal circuits with small numbers of identifiable cells are ideal preparations for investigating the mechanistic basis of modulation and plasticity. Such studies have, however, been limited by the lack of pharmacological tools. Although the protein sequences and second-messenger couplings of 5-HT receptors are relatively well conserved across all species,their pharmacological profiles can vary significantly between vertebrate and invertebrate preparations, and even within invertebrate preparations(Tierney, 2001). This is not surprising as one would not expect selection for or against amino acids involved in binding synthetic ligands. We have recently identified and cloned two crustacean 5-HT receptors from the spiny lobster, Panulirus interruptus, infraorder Achelata: 5-HT_1αPan and 5-HT_2βPan (Clark et al.,2004; Sosa et al.,2004). We were interested to know whether the actions of pharmacological agents were conserved across crustacean species. We therefore obtained full-length clones for the same two receptors from a second distantly related decapod crustacean, the swamp crayfish, Procambarus clarkii(infraorder Astacidea), and expressed these receptors in a heterologous system in order to determine their second messenger couplings and pharmacological profiles. These parameters were then compared across the two species.

Crayfish, Procambarus clarkii Girard 1852, were obtained from Atchafalaya Biological Supply (Raceland, LA, USA) and kept communally at 20°C in 40 l tanks with continual filtration and aeration.

HEK293, NIH3T3, COS-7 and MDCK cells, Earle's minimal essential medium(EMEM), horse serum, trypsin, penicillin and streptomycin were obtained from the American Type Culture Collection (Mannassas, VA, USA). Dulbecco's modified Eagle's medium (DMEM) was from Mediatech Inc. (Herndon, VA, USA). Dialyzed fetal bovine serum (FBS), TRex cell line (293-TR), pDNA4/TO plasmid,blasticidin and zeocin were from Invitrogen (Carlsbad, CA, USA). Cinanserin was obtained from Tocris (Ballwin, MO, USA). All other chemicals were from Sigma (St Louis, MO, USA). For pharmacology experiments, amine and agonist stock solutions (10^–1 mol l^–1) were made fresh for every experiment in medium or 50% ethanol, respectively. Two exceptions were tyramine (Tyr), which was made fresh as a 10^–2 mol l^–1 stock in medium, and methysergide, which was made as a 10^–2 mol l^–1 stock in DMSO and stored at –20°C. Antagonist drugs were made as 10^–2 mol l^–1 stock solutions in DMSO and stored at –20°C.

Complete cloning and sequencing of 5-HT_2βPan and 5-HT_1αPan from Panulirus interruptus have been previously described (Clark et al.,2004; Sosa et al.,2004).

We also previously cloned a large segment of 5-HT_1αProspanning transmembrane domains III–VII from Procambarus clarkii(Sosa et al., 2004). We have now completed the sequencing of the 5-HT_1αPro cDNA using rapid amplification of DNA ends (SMART RACE cDNA Amplification kit; BD Biosciences, Clontech, Cambridge, UK) as previously described(Clark et al., 2004). Constructs containing the complete ORF were assembled using standard procedures (Ausubel et al.,1990). Both strands of the construct were sequenced and errors that had been introduced in the cloning process were corrected using QuikChange site directed mutagenesis (Stratagene, La Jolla, CA, USA). The construct was then cloned into the pDNA4/TO (Invitrogen) expression plasmid.

5-HT_2βPro was cloned from crayfish cDNA using degenerate RT-PCR and RACE. Previously, 5-HT_2βPan had been identified in the Drosophila genome database and the ortholog from Panulirus was fully cloned and characterized for signal transduction properties (Clark et al.,2004). Degenerate primers were designed based on conserved regions of these Panulirus and Drosophila orthologs of 5-HT_2βPro (written 5′-3′): 5-5-1,GAYGTIYTITTYTGYACIGCIWSIATHATG; 5-5-2, ATGCAYYTITGYACIYTIWSIGTIGAYMGI TT; 5-3,CATDATDATIARIGGDATRTARAARCA; 3-5-1, CAYGGIMGIAAYATHMGIATGGARCA; 3-5-2,WUIGARCARAARGCNACNAARGU; 3-3, YUURUURAAIAWIGURUARRA.

Multiple cDNA preparations, each from a separate crayfish nervous tissue mRNA preparation, were used as templates for nested PCR experiments with these degenerate primers to amplify fragments of the crayfish ortholog, as described previously (Baro et al., 1994; Sosa et al., 2004). Primers specific to 5-HT_2βPro were designed to generate a large clone of 5-HT_2βPro. The N- and C-terminals of 5-HT_2βPro were then cloned using SMART RACE as described above. A construct containing the complete ORF was assembled, sequenced and inserted into the pIRESneo (Clontech, Mountain View, CA, USA) expression plasmid as previously described for 5-HT_2βPan(Clark et al., 2004).

Sequence data were analyzed using Sequencher 4.1 (Gene Codes Corp., Ann Arbor, MI, USA). Sequences for other arthropod species were obtained from GenBank and alignments were created to determine sequence identities using the ClustalW algorithm with default settings in Lasergene MegAlign (DNASTAR Inc.,Madison, WI, USA). Full Procambarus sequences have been deposited in GenBank under accession numbers EU131666 (5-HT_2βPro) and EU131667 (5-HT_1αPro).

We previously characterized the signaling and pharmacological properties of the 5-HT_2βPan receptor transiently expressed in HEK293 cells(Clark et al., 2004). We used the same techniques to characterize the Procambarus ortholog,5-HT_2βPro. Briefly, cells were maintained in EMEM supplemented with 10% FBS, 50 i.u. ml^–1 penicillin and 50 μg ml^–1 streptomycin (normal medium). Cells were plated on 60 mm dishes in EMEM without antibiotics, allowed to grow to 95–100%confluency and then transfected with 2 μg of DNA using lipofectamine(Invitrogen). The plates were supplemented to a final concentration of 10% FBS 6 h after transfection, and the medium was replaced with normal medium 24 h after transfection.

In order to functionally characterize 5-HT_1αPan and 5-HT_1αPro, we first generated full-length constructs using standard recombinant techniques, as described above. The 5-HT_1α receptors were first cloned into pIRESneo and stably transfected into several cell lines using lipofectamine as detailed in the Results. Immunoblotting experiments coupled with the lack of growth after several weeks of selection suggested that none of these cell lines could stably express either the 5-HT_1αPan or the 5-HT_1αPro receptor. This did not appear to be a general phenomenon associated with crustacean receptors, as we successfully expressed 5HT_2β receptors and crustacean dopamine receptors in these cell lines. The most reasonable explanation for our finding was that the cell lines could not tolerate high levels of the 5-HT_1α receptor,and for reasons unknown, cells expressing 5-HT_1α receptors were selected against. Inducible expression systems have been developed to deal with this type of problem.

We next expressed the 5-HT_1α constructs in an inducible expression system. In this system the 5-HT_1α cDNA is stably integrated into the genome of the parental HEK cells, but it is not transcribed unless tetracycline is present in the medium. Using this system we could express 5-HT_1α receptors for a defined time interval,and then perform the assays before the cells were selected against. 5-HT_1α cDNA was cloned into the inducible expression plasmid,pDNA4/TO, behind the tetracycline operator. The new constructs were then transfected into HEK293-TR cells stably expressing the tetracycline repressor protein. Stably transfected cells (i.e. those in which the plasmid carrying the receptor was incorporated into the HEK cell genome) were selected for>4 weeks in DMEM supplemented with 10% dialyzed fetal bovine serum, 5 μg ml^–1 blasticidin and 300 μg ml^–1 zeocin(complete medium; TRex regulated expression system, Invitrogen). Confluent plates were easily obtained after 4 weeks, suggesting that the 5-HT_1α cDNA could be maintained in tissue culture cells as long as it was not transcribed and translated. Next, western blots were used to confirm tet-repressor-regulated 5-HT_1αPan expression as follows. Cells were plated in 60 mm dishes and induced with 1 μmol l^–1 tetracycline 0, 6, 8, 12 and 24 h before collection and isolation of protein by previously described methods(Clark et al., 2004; Sosa et al., 2004). Protein preparations were run on a 10% SDS-PAGE gel, transferred to PVDF membranes and probed with a custom-made rabbit anti-5-HT_1αCrust antibody(Sosa et al., 2004). Bands were visualized using chemiluminescence (Immun-Star, BioRad, Hercules, CA,USA). 5-HT_1αPan production commences within 6 h after induction and lasts through 24 h of induction(Fig. 1). Crude protein prepared from lobster nervous system was run as a positive control. Similar results were found for induction of 5-HT_1αPro in stably transfected 293-TR-5-HT_1αPro lines. Based on these findings we induced cells for functional assays for 18–20 h (below).

Inositol phosphate (IP) release was assayed as previously described(Clark et al., 2004). Briefly,transiently transfected cells were divided among wells on a 24-well plate with 1 μCi ml^–1 of [³H]myoinositol (Amersham,Piscataway, NJ, USA) and allowed to grow to 95–100% confluency over 48 h. The cells were washed with fresh EMEM and then exposed to 10 mmol l^–1 LiCl in EMEM for 20 min at 37°C. As applicable,antagonists were added to individual wells and allowed to incubate for an additional 10 min. 5-HT or agonist drugs were added to test well contents and cells were returned to 37°C for 60 min. The medium was removed and replaced with ice-cold 20 mmol l^–1 formic acid. Plates were then placed on ice for 30 min. The cell lysate was collected and applied to AG1-X8 columns (BioRad, Hercules, CA, USA) equilibrated with 20 mmol l^–1 formic acid. The columns were washed with 50 mmol l^–1 ammonium hydroxide followed by elution of inositol phosphates (IP) with 10 ml of 1 mol l^–1 ammonium formate–0.1 mol l^–1 formic acid. The IP fraction was scintillation counted. Membranes attached to the wells were dissolved in 1 mol l^–1 NaOH and scintillation counted as total phosphatidyl inositols (PI). Activation results are expressed as the fraction of radioactivity incorporated in IP over that in IP+PI and normalized to activity observed in negative control wells (no drug) for every experiment.

Cyclic AMP levels in HEK293-TR cells stably expressing 5-HT_1αPan or 5-HT_1αPro were determined using a Direct cAMP kit (Assay Designs, Ann Arbor, MI, USA) as previously described(Clark et al., 2004). For receptor activation assays, stably transfected cells were plated in 24-well plates and allowed to grow to 100% confluency. The medium was replaced with 1 ml of complete medium containing 1 μg ml^–1 tetracycline to induce expression of receptor protein. After 18–20 h the medium was replaced with 1 ml of fresh DMEM containing 2.5 mmol l^–13-isobutyl-1-methylxanthine to block phosphodiesterase activity and plates were incubated for 10 min. Antagonists were added to individual wells (if applicable) and allowed to incubate for an additional 10 min. 5-HT or agonists and forskolin (250 nmol l^–1), a nonspecific activator of adenylyl cyclase, were then added to individual wells and left at 37°C for 30 min. The medium was removed and replaced with 0.5 ml of 0.1 mol l^–1 HCl containing 0.8% Triton X-100. Plates were shaken for 30 min at room temperature, the lysate collected, centrifuged for 5 min at 600 g and the supernatant assayed for [cAMP] using the Direct cAMP kit and [protein] using a BCA Protein Assay Kit (Pierce, Rockford, IL, USA). Data are presented as pmoles of cAMP per milligram of protein.

Data for all pharmacology assays involving the heterologous expression systems were plotted and analyzed in GraphPad Prism v.4. Statistics for bar graphs were calculated using a two-way ANOVA with a Bonferroni post-test. Dose–response curves were fitted with a standard slope top–bottom or bottom–top dose–response curve to calculate EC/IC₅₀and efficacy values.

The crustacean 5-HT_2β receptor was recently cloned from Panulirus interruptus and characterized with respect to its signaling properties (Clark et al.,2004) and pharmacological profile (N.S. and D.J.B., unpublished). We have used degenerate RT-PCR and RACE to clone the Procambarus clarkii ortholog, 5-HT_2βPro, from crayfish nervous system cDNA. The full-length 5-HT_1α cDNA was previously cloned from Panulirus and the partial sequence was reported for Macrobrachium rosenbergii and Procambarus clarkii(Sosa et al., 2004). In this study we completed the cloning and sequencing of the Procambarus5-HT_1α ortholog using RACE. The predicted amino acid sequences of 5-HT_2β and 5-HT_1α orthologs were very well conserved between Panulirus and Procambarus; 72 and 90%, respectively, over the entire protein (Figs 2 and 3, Table 1).

When comparing 5-HT receptors from various species, the overall sequence identity can fall very quickly. This is mainly due to the N- and C-terminal domains and the majority of the third intracellular loop, all of which are highly variable in 5-HT receptors. These regions are not thought to be critical to signaling or pharmacology(Kroeze et al., 2002; Saudou et al., 1992; Witz et al., 1990). When these variable regions were excluded from the alignment, a core region representing 34–59% of the protein remained. This core consisted of transmembrane domains and short linker regions important for maintenance of protein structure, ligand binding and signaling. The identity between Panulirus and Procambarus orthologs in the core protein was very high at 97% for 5-HT_2β and 98% for 5-HT_1α (Table 1). The complete 5-HT_2β sequence from each crustacean was 45% identical to the predicted protein sequence of their ortholog from the fruit fly, with an increase to 68% for the core protein. Similarly, crustacean 5-HT_1α receptors showed 29–53%identity to orthologs from fly, budworm and butterfly with the core protein sharing at least 76% identity between any of these five arthropod species. In addition, the cores of both crustacean receptors had greater than 40% identity to their human homologs.

5-HT receptors from all species share key conserved residues in their transmembrane domains that are responsible for forming the ligand binding pocket. Because the N-terminal and extra-membrane loops of diverse GPCRs are of variable lengths, referring to residues by their absolute position within these proteins does not allow for comparisons of structural domains between proteins. The Ballesteros-Weinstein numbering scheme used here identified a crucial and conserved characteristic residue common to all G protein coupled receptors (GPCRs) within each transmembrane domain (TM) that was arbitrarily assigned the number 50 (grey, Figs 2 and 3). Other residues within that TM domain were then numbered with the TM number followed by the residue number in relation to this identified amino acid [i.e. Phe (6.51) immediately follows the reference residue in TM6, Pro (6.50)](Ballesteros and Weinstein,1995). In biogenic amine receptors the charged residue Asp (3.32)is thought to act as a counterion for the protonated amine moiety of amine ligands, agonists and antagonists and is required for ligand binding but not receptor activation (Kristiansen et al.,2000). The presence of Asp (3.32) in combination with Trp (7.40)is considered a unique fingerprint for biogenic and trace amine GPCRs. These amino acids are involved in ligand binding and receptor activation. In addition to these residues, 5-HT receptors typically have a conserved group of hydrophobic amino acids [Trp (3.28), Phe (5.47), Phe (5.48), Trp (6.48), Phe(6.51), Phe (6.52), Trp (7.40), Tyr (7.43)] that form the hydrophobic ligand-binding pocket within the tertiary structure of the receptor[Kristiansen (Kristiansen,2004) and references therein;(Roth et al., 1997)]. The hydroxyl group of 5-HT is thought to be stabilized by the Ser (5.43) residue in transmembrane helix 5. Finally, a disulfide bridge formed between Cys(3.25) and a Cys in extracellular loop 2 (EL 2) is important in maintaining tertiary structure and stabilizing the ligand binding pocket. All of these key amino acids (first defined in mammalian receptors) were conserved in 5-HT_2β and 5-HT_1α from crayfish and spiny lobster (Figs 2 and 3).

Crucial residues for Gα subunit binding specificity are located in cytoplasmic amphipathic α-helical extensions of TM5 and 6 in intracellular loop 3 (IL3) (reviewed by Blenau and Baumann, 2001; Gether, 2000; Kristiansen, 2004; Strader et al., 1995). The C-terminal region of IL3 near the membrane interface with TM6 interacts with the C-terminal end of the Gα protein, helping to confer the receptor's specificity for a specific Gα subtype. Again, these regions were very well conserved when comparing the crustacean receptor orthologs to each other,to orthologs from other arthropods and to a human homolog (Figs 2 and 3). Although IL2 does not appear to be involved in determining Gα specificity, this loop contains an α-helix that is thought to act in conjunction with the adjacent DRY motif as a switch between active and inactive states making this loop important for efficient G protein activation. Ligand binding results in protonation of the Asp (3.49) in the DRY motif and in significant rotation of TM3 and TM6 and transition to the active state of the receptor. In addition,the DRY motif and the residues surrounding it are important for constitutive activation in 5-HT receptors (Gether,2000; Shapiro et al.,2002). Interestingly, the 5-HT_2β receptor cloned from Panulirus has evolved a DRF sequence in place of the DRY that confers increased basal activity of the receptor when stably expressed in cell culture (Clark et al., 2004);this sequence alteration was conserved in 5-HT_2βPro from crayfish (Fig. 2).

Serotonin receptors are extensively post-translationally modified by several mechanisms. Most known GPCRs, including 5-HT_2βPro and 5-HT_1αPro have several consensus sites for N-linked glycosylation (Asn–X–Ser/Thr) in the N-terminal tail (Figs 2 and 3) and sometimes in other extracellular regions such as EL2. Proper glycosylation of at least some of these sites is required to obtain appropriate levels of receptor expression on the cell surface (Lanctot et al.,2006). Efficiency of ligand binding and functional activity of receptors are not known to be affected by the glycosylation state in receptors that are expressed in the membrane. The putative glycosylation sites of 5-HT_2β and 5-HT_1α were conserved between orthologs from Panulirus and Procambarus (Figs 2 and 3). Many GPCRs also have a Cys residue in the proximal region of the C-terminal tail that is a putative palmitoylation site; this creates a membrane anchor, generating an additional cytoplasmic loop. This site was present in crustacean 5-HT_2βreceptors but not in 5-HT_1α.

The high degree of conservation of the key structures within the crustacean receptors led us to predict that their signaling pathways will also be the same as those of their vertebrate and invertebrate homologs. Because of the high level of overall conservation, we also might expect the spiny lobster and crayfish orthologs of 5-HT_2β and 5-HT_1α to exhibit similar pharmacological profiles. In order to compare the functional properties of Panulirus and Procambarus 5-HT receptors, we heterologously expressed the proteins in cell culture and used second messenger assays to determine their ligand specificity, signaling and pharmacological properties.

The arthropod 5-HT_2β receptor was initially cloned from Panulirus and shown to be specific for 5-HT over other biogenic amines (Clark et al., 2004). When stably expressed in HEK cells, activation of this receptor resulted in increased intracellular levels of inositol phosphates (IP), activation of protein kinase C (PKC) and no change in cAMP levels. In addition, stably expressed 5-HT_2βPan demonstrated constitutive activity conferred by the DRY motif (Clark et al.,2004). In this study we transiently expressed 5-HT_2βPan in HEK cells and measured IP release in response to amines and putative pharmacological agents. Interestingly, unlike stably transfected cultures, we found no constitutive activity of 5-HT_2βPan when the receptor was transiently expressed (see Discussion).

Non-transfected parental HEK cells did not respond to 1 mmol l^–1 concentrations of any of the monoamines in the IP assay(Fig. 4A). 5-HT_2βPan responded to 5-HT, dopamine and tyramine with IP release (Fig. 4B). The EC₅₀ for 5-HT was 52 nmol l^–1 whereas greater than 50 μmol l^–1 dopamine (DA) and tyramine (Tyr) were required to activate 5-HT_2βPan (Fig. 4C). At 1 mmol l^–1 these amines also had an efficacy less than 55% that of 5-HT (Table 2), indicating that 5-HT is the preferred functional ligand for the 5-HT_2βPan receptor. As observed for transiently expressed 5-HT_2βPan, we found no constitutive activity of transiently expressed 5-HT_2βPro. 5-HT_2βPro responded strongly to 5-HT with a smaller response to DA(Fig. 4D). The EC₅₀for 5-HT_2βPro was 270 nmol l^–1 whereas 1 mmol l^–1 DA elicited only a minimal response(Fig. 4E, Table 2).

We were not able to express 5-HT_1αPan or 5-HT_1αPro in traditional systems including HEK293, NIH3T3,MDCK or COS-7 cells; all cells that produced the receptor protein, as determined by western blot analysis, were unhealthy and did not grow beyond 3 weeks. In the few cases where a stable cell line was generated,5-HT_1α protein could not be detected by western blot analysis, suggesting rearrangements in the DNA construct. We do not understand why mammalian cell lines were unable to stably express 5-HT_1αusing traditional methods. 5-HT_1αPan and 5-HT_1αPro appeared to be constitutively active (below) so high levels of expression of the receptors in standard expression systems may have resulted in toxicity. Alternatively, the protein synthesis, export or turnover machineries of the cells may have been overly taxed by high levels of 5-HT_1α expression such that they were not able to maintain normal functions.

In order to functionally characterize 5-HT_1αPan and 5-HT_1αPro, we therefore employed an inducible expression system. 5-HT_1αPan or 5-HT_1αPro constructs were stably transfected into 293-TR cells expressing the tetracycline repressor protein. In this system the 5-HT_1α construct is under control of the Tet operator sequence, which binds the repressor protein in the absence of tetracycline, thereby preventing expression of 5-HT_1α. Upon addition of tetracycline to the media, the repressor protein dissociates and 5-HT_1α is transcribed and translated into protein. The western blot in Fig. 1 indicates that non-induced cells (tetracycline absent) did not express detectable levels of 5-HT_1αPan. After induction (tetracycline present) we were able to obtain high levels of 5-HT_1αPan expression within 6 h that lasted for at least 24 h (Fig. 1). Similar results were obtained for cells induced to express 5-HT_1αPro (not shown).

When compared with non-induced cells, induced cells expressing either 5-HT_1αPan or 5-HT_1αPro showed an increased sensitivity to forskolin, a non-specific activator of adenylyl cyclase(Fig. 5A,C). This supersensitization of adenylyl cyclase is typical of cells expressing a constitutively active Gi/o-coupled receptor(Johnston and Watts, 2003). Constitutive activity has been observed in mammalian 5-HT₁receptors as well as other G protein-coupled receptors(Berg et al., 2005; Cosi and Koek, 2000; Johnston and Watts, 2003; Liu et al., 1999).

All known vertebrate and invertebrate 5-HT₁ receptors inhibit adenylyl cyclase, resulting in decreased cAMP levels after stimulation with forskolin (Hoyer et al., 2002; Tierney, 2001). As expected for a 5-HT₁ receptor expressed in a HEK293 cell line,5-HT_1αPan activation with 5-HT inhibited forskolin-stimulated cAMP accumulation, presumably via Gi/o inhibition of adenylyl cyclase. At 1 mmol l^–1 concentrations, 5-HT was the only monoamine to significantly activate 5-HT_1αPan(Fig. 5A). 5-HT is a highly effective ligand at 5-HT_1αPan with an EC₅₀ of 8.4 nmol l^–1 (Fig. 5B, Table 2). No significant change in cAMP levels was observed with any biogenic amine in non-induced 293-TR-5-HT_1αPan cells(Fig. 5A).

Similarly, 5-HT_1αPro activation with 1 mmol l^–1 5-HT also blocks forskolin-stimulated cAMP accumulation. Tyr was the only other biogenic amine that resulted in a significant decrease of cAMP in cells expressing 5-HT_1αPro(Fig. 5C). Serotonin and tyramine are therefore the only amines that significantly inhibit adenylyl cyclase via activation of the 5-HT_1αPro receptor. Histamine, dopamine and octopamine all produced an increase in cAMP levels in non-induced cells. Similar levels of cAMP were observed in the induced cells,suggesting that these amines do not act at the 5-HT_1αProreceptor. Non-induced 293-TR-5-HT_1αPro cells gave a positive cAMP response at high 5-HT concentrations(Fig. 5D, crosses). Unlike the cell line expressing Panulirus 5-HT_1α, during the selection period, the 5-HT_1αPro cell line appears to have initiated expression of an endogenous 5-HT receptor that is positively coupled to adenylyl cyclase. It is not unusual for cell cultures to change their karyotypes or expression profiles over time due to the lack of selection pressure to maintain a constant genome. Such changes can be significant and can lead to different net signaling effects for the same protein(Clark and Baro, 2007; Friedman et al., 2002). Cells that were induced to produce 5-HT_1αPro responded to low 5-HT concentrations with a decrease in forskolin-induced cAMP mediated by 5-HT_1αPro. This effect was dampened and partially reversed at higher 5-HT concentrations through activation of the endogenous receptor(Fig. 5D, black squares). Subtraction of the 5-HT curves of non-induced cells(Fig. 5D, crosses) from induced cells (Fig. 5D, black squares)resulted in a sigmoidal dose–response for 5-HT at 5-HT_1αPro with an EC₅₀ of 31 nmol l^–1 (Fig. 5D,grey squares). Tyramine was an inefficient agonist and slightly reduced cAMP accumulation only at high concentrations(Fig. 5D, triangles) indicating that 5-HT is the preferred functional ligand for 5-HT_1αPro.

In order to determine if the conservation in sequence and signaling extends to the receptors' responses to pharmacological agents, we tested crayfish 5-HT_2βPro and 5-HT_1αPro with a suite of agonist drugs. All drugs that showed significant activity at 10^–3 mol l^–1 in an initial overview(Fig. 6) were tested in dose–response curves (see Fig. 7 for examples), which were then used to determine the potency and efficacy (% of maximum 5-HT response at highest concentration) of the drugs. The EC₅₀ is a measure of the potency of a drug and reflects its binding affinity at the receptor. Because the maximum effect, or efficacy,achieved by any drug is dependent on the number of receptors expressed we ran a parallel dose–response curve for 5-HT in every experiment and normalized all agonist maximum effects to the maximum 5-HT response, set at 100%. These data have been previously reported for the Panulirusreceptor (N.S. and D.J.B., unpublished) and are summarized with the Procambarus receptor data in Table 2 for comparison.

Most of the drugs that activated 5-HT_2βPro had an efficacy close to that of 5-HT (Fig. 6A, Table 2). Methysergide(EC₅₀=110 nmol l^–1) was more potent than 5-HT(EC₅₀=270 nmol l^–1), however, it elicited only 19%of the total response obtained with 5-HT. Thus, when both potency and efficacy are considered, 5-HT is the strongest ligand at 5-HT_2βPro. The rank potency of effective agonists at 5-HT_2βPro was methysergide>5-HT>8-OH-DPAT>MeOTryp>5-CT>2-Me-5-HT>α-Me-5-HT. This was very similar to the potency ranking of these drugs at 5-HT_2βPan. However, as can be seen in Table 2, the potency of drugs at the crayfish receptor was generally lower than in lobster, with 5-HT being five times less potent. The crayfish ortholog did not respond to 2,5-dimethoxy-4-iodoamphetamine (DOI), which was a relatively weak agonist of 5-HT_2βPan. The potencies of methysergide,5-carboxamidotryptamine (5-CT) and 5-methoxytryptamine (MeOTryp) were similar for the two orthologs, but their efficacies were twofold lower at the 5-HT_2βPro ortholog (Table 2). However, 2-methyl-serotonin (2-Me-5-HT) was less potent at 5-HT_2βPro, whereas ortholog efficacies were equivalent. Together these data indicate that, whereas the pharmacological profiles of 5-HT_2βPro and 5-HT_2βPan are conserved in terms of which drugs are active, 5-HT_2βPro was consistently less sensitive to agonist stimulation. As observed for 5-HT_2βPan,no change in IP level was detected in cells expressing 5-HT_2βPro after application of 10^–3 mol l^–1N-acetyl-5-HT, quipazine, or 1-(m-chlorophenyl)-piperazine (mCPP). None of the drugs had significant effects on non-transfected parental HEK cells(Fig. 6A).

Most of the putative agonists tested (at 1 mmol l^–1)resulted in some activation of the 5-HT_1αPro receptor(Fig. 6B, Table 2). The rank potency of effective agonists at 5-HT_1αPro was 5-HT > 2-Me-5-HT >α-Me-5-HT > methysergide > tyramine > 8-OH-DPAT > mCPP >quipazine. The agonist 2-Me-5-HT was almost as potent but only 73% as efficacious as 5-HT. This pharmacological profile is very similar to that of its lobster ortholog, 5-HT_1αPan. However, as for the 5-HT_2βPan receptor, the crayfish ortholog of 5-HT_1αPro was generally less sensitive to agonists than the lobster ortholog. ±-Methyl-serotonin (α-Me-5-HT) was more potent,and (±)-8-hydroxy-2-(di-n-dipropylamino) tetralin (8-OH-DPAT)was less potent at 5-HT_1αPro relative to 5-HT_1αPan, however, both these drugs had lower efficacies at the crayfish ortholog. The potency of mCPP was comparable for Panulirus and Procambarus 5-HT_1α, although it was less efficient at the crayfish ortholog. Methysergide was less potent but had a higher efficacy at 5-HT_1αPro compared to 5-HT_1αPan. Two drugs that could not be tested at 5-HT_1αPan because they altered cAMP levels in non-induced 293-TR-5-HT_1αPan cells, 2-Me-5-HT and quipazine, are effective agonists of 5-HT_1αPro. Several drugs (DOI, 5-CT,MeOTryp, N-acetyl-5-HT) had complex effects on induced 293-TR-5-HT_1αPro cells that could not be fitted with standard dose–response curves. These complex effects are probably due to endogenous 5-HT receptors expressed by the non-induced cell line as observed in the 5-HT dose–response curve above(Fig. 5D, crosses). We were therefore not able to determine an EC₅₀ or relative efficacy measurements for these drugs.

In summary, we identified two agonists that would activate 5-HT_1α but not 5-HT_2β(Table 2). mCPP is an agonist of 5-HT_1α but not 5-HT_2β in Procambarus and Panulirus whereas quipazine is also inactive at 5-HT_2β from either species and activates 5-HT_1αPro but could not be tested on 5-HT_1αPan.

Because pharmacological agents can be active at multiple 5-HT receptors,strategic combinations of drugs will be necessary to identify the receptors involved in physiological 5-HT effects. In addition to agonists, we therefore tested a suite of putative antagonists on 5-HT_2β and 5-HT_1α from Procambarus and compared them to the antagonist profile for Panulirus receptors (N.S. and D.J.B.,unpublished).

Antagonists were applied to cells 10 min before 5-HT application and second messenger assays were used to test receptor activation. Antagonists were first screened at 10^–5 mol l^–1(Fig. 8) and then dose–response curves were generated for any drugs that significantly blocked 5-HT activation of second messengers at that concentration (see Fig. 7 for examples). The IC₅₀ was calculated and is reported as a measure of potency for the drug. The efficacy, or maximum effect, for each drug again depends on receptor expression levels and is therefore reported as a percentage reduction from the level of receptor activation achieved by 5-HT alone in the same experiment(Table 3).

Putative antagonists had no effect on parental HEK cells(Fig. 8A). Several antagonists(at 10 μmol l^–1) significantly reduced 5-HT-induced (1 mmol l^–1) IP release in HEK cells expressing 5-HT_2βPro (Fig. 8A, Table 3). The rank potency of effective antagonists at 5-HT_2βPro was(+)butaclamol > methiothepin > ritanserin > cinanserin > clozapine(Table 3). Of the antagonists tested, ketanserin, spiperone, prazosin, (–)butaclamol, gramine and atropine had no effect at 10^–5 mol l^–1. The potency profile of active antagonists was nearly conserved between 5-HT_2β from crayfish and lobster, but four of the five drugs had lower IC₅₀s at the crayfish ortholog. In addition, all of the drugs blocked the crayfish receptor by >80%, but the lobster receptor by only 48–84%. 5-HT_2β from lobster and crayfish diverge by almost 30% in the variable regions (Table 1) and these differences could contribute to the increase in potency and efficacy of antagonists at 5-HT_2βPro.

Twenty-nine putative antagonists, including those that were characterized at 5-HT_2βPro, were tested on 5-HT_1αPro(Fig. 8B, Table 3). None of 14 putative antagonists (at 10 μmol l^–1) were able to block 5-HT (at 10 μmol l^–1) activation of 5-HT_1αPro. The antagonists were ineffective even at 10 nmol l^–1 (data not shown). These results agree with what we found for 5-HT_1αPanfrom Panulirus where the same putative antagonists also did not block activation of the receptor (Table 3). Because of background activity, presumably at endogenous GPCRs in the non-induced cell line, we were not able to test the 15 other putative antagonists on cells induced to express 5-HT_1αPro(Table 3). It is not surprising that the uninduced Pan and Pro cell lines showed different background activities, as it is well established that cell lines diverge with time in culture (Clark and Baro,2007). Further, multiple copies of the plasmid DNA can be integrated into the host genome, and this has the potential to alter expression of nearby genes. It is unlikely that the number or location of plasmid insertions would be identical in the two cell lines.

As was seen at 5-HT_1αPan, several putative antagonists appeared to increase the efficacy of the 5-HT effect on cAMP accumulation in cells expressing 5-HT_1αPro(Fig. 8B). When tested without 5-HT, however, these drugs had no significant effect on forskolin-stimulated cAMP production (not shown), indicating that they are not acting as agonists. Further studies will be necessary to determine the relationship between antagonist drugs and the crustacean 5-HT_1α receptors.

Although we could not find effective 5-HT_1αProantagonists, we were able to identify several drugs that would block 5-HT_2β while not affecting 5-HT activation of 5-HT_1α (Table 3). Methiothepin and cinanserin efficiently blocked 5-HT_2βPro but had no activity at 5-HT_1αPro. Similarly, cinanserin, (+)butaclamol and ritanserin all blocked 5-HT_2βPan but not 5-HT_1αPan. All four of these drugs are effective 5-HT_2β antagonists but some could not be tested at the 5-HT_1α receptor from one or the other species to confirm specificity because of differences in the parental 5-HT_1α cell lines.

Two 5-HT receptors from two infraorders of decapod crustaceans, Panulirus (Achelata) and Procambarus (Astacidea) are highly conserved in their sequence, signaling and pharmacological properties. A significant portion of the 5-HT_1α receptor from the giant freshwater prawn Macrobrachium rosenbergii, has also been cloned and is also highly conserved with its lobster and crayfish orthologs(Sosa et al., 2004). Together,these data suggest that the conservation of pharmacological function observed between the species studied here may extend beyond reptantian crustaceans, in which case the drugs identified here could perhaps be useful for dissecting 5-HT mechanisms in diverse crustacean systems. We have identified several antagonists, (+)butaclamol, cinanserin and ritanserin for Panulirusand methiothepin and cinanserin for Procambarus, that will block 5-HT_2β while leaving 5-HT_1α unaffected. In addition, mCPP weakly activates 5-HT_1α while having no effect on 5-HT_2β for both species whereas quipazine activates 5-HT_1αPro but not 5-HT_2βPro. Combinations of these drugs can be applied in studies of the mechanisms underlying 5-HT modulation in identified circuits and cells of the California spiny lobster,crayfish and related crustacean nervous systems.

Panulirus and Procambarus orthologs of 5-HT_2β and 5-HT_1α have the 5-HT signature sequences required for ligand binding and G protein coupling. Most of the sequence differences between receptor orthologs were located in the amino termini and the nonconserved center of the third intracellular loop. These regions may contribute to differences in sensitivity to pharmacological agents observed between receptor orthologs; however, no specific function has been ascribed to these regions to date (Kroeze et al., 2002). Indeed, in characterizing the Drosophila5-HT₁ and 5-HT₇ receptors, the amino termini were removed to increase expression levels in cell culture, with no apparent effect on receptor function (Saudou et al.,1992; Witz et al.,1990).

5-HT was the only biogenic amine to significantly activate 5-HT_2βPro, however, DA and Tyr were also weak agonists of 5-HT_2βPan, suggesting that these amines could activate 5-HT_2βPan in a physiological context. By contrast, Clark et al. (Clark et al., 2004) found that 5-HT was the only biogenic amine to significantly activate 5-HT_2βPan. Thus, it is not clear whether or not DA and Tyr have any effect on the 5-HT_2βPan receptor in the native system. If these amines do activate native receptors, then is likely to be a concern only at synaptic sites where concentrations can reach up to 1 mmol l^–1 [(Clements,1996; Frerking and Wilson,1996) and references therein] because, high levels of DA and Tyr are required to activate 5-HT_2βPan. The difference between the two studies may be because we measured IP release [i.e. phospholipase Cβ(PLCβ) activity] whereas the previous study measured PKC activity. Because PKC is downstream of PLCβ and dependent on the consequent release of Ca²⁺, low levels of PLCβ activation by DA and Tyr may not be sufficient to initiate a cascade culminating in PKC activation. Alternatively, the difference may stem from our use of transiently transfected cells whereas Clark et al. (Clark et al.,2004) studied stable transfectants. In addition,5-HT_2βPan was found to be constitutively active when stably expressed in HEK293 cells (Clark et al.,2004). However, in our transient transfections with 5-HT_2βPan and 5-HT_2βPro we found no constitutive activity. The reason for the difference is unclear.

Crustacean 5-HT_1α receptors are typical type 1 5-HT receptors, preferentially responding to 5-HT over other biogenic amines with EC₅₀ values of 8.4 and 31 nmol l^–1, respectively. Other arthropod orthologs of 5-HT_1αPan and 5-HT_1αPro have been cloned and characterized from Drosophila (5-HT_1ADro, originally 5-HT_dro2A)and Boophilus microplus and these have comparable EC₅₀values of 30 and 83 nmol l^–1 for 5-HT, respectively(Chen et al., 2004; Saudou et al., 1992). Whereas the lobster ortholog responded only to 5-HT, the crayfish ortholog was also activated by Tyr. Activation of 5-HT₁ receptors, including those described here, results in inhibition of forskolin-stimulated cAMP accumulation. In addition, the crustacean 5-HT_1α receptor shows agonist-independent activity when expressed in HEK cells.

The pharmacological profiles of 5-HT receptors are very similar for Procambarus and Panulirus, but show more variability when compared with homologs from other invertebrates. Agonist activity at crustacean 5-HT_2β and 5-HT_1α orthologs is quite conserved. Two agonists that would differentiate between 5-HT_2β and 5-HT_1α were identified. mCPP activates 5-HT_1α but not 5-HT_2β in both Procambarus and Panulirus. Quipazine is also specific to 5-HT_1αPro over 5-HT_2βPro and, similarly, is inactive at 5-HT_2βPan but could not be tested on 5-HT_1αPan. Quipazine binds weakly to a molluskan 5-HT₁ receptor (Sugamori et al., 1993) and to 5-HT_2αDro(Colas et al., 1995),indicating that its specificity may not be highly conserved among invertebrates. 8-OH-DPAT was an effective agonist of both crustacean 5-HT_2β and 5-HT_1α, and is also the only agonist reported to bind 5-HT_1αDro(Saudou et al., 1992). 8-OH-DPAT was considered a specific mammalian 5-HT₁ agonist but was later found to also activate 5-HT₇ receptors(Bard et al., 1993; Sprouse et al., 2004) and it activates Drosophila 5-HT_7Dro(Witz et al., 1990). Interestingly, methysergide is an agonist at crustacean 5-HT_2βand 5-HT_1α. This drug is a functional antagonist of 5-HT_7Dro (Witz et al.,1990) and of vertebrate 5-HT₂ receptors but has agonist activity at some vertebrate 5-HT₁ receptors(Silberstein, 1998).

The antagonist profiles of 5-HT_2β receptors from Panulirus and Procambarus are also very well conserved. Cinanserin, (+)butaclamol, ritanserin and methiothepin were identified as antagonists that would block crustacean 5-HT_2β but not 5-HT_1α receptors, however not all of these could be tested at 5-HT_1α from both species. The antagonist profiles of the crustacean receptors could not be compared to other arthropod orthologs, as 5-HT_2β receptors have only been cloned from crustaceans,though they have been found in the fruit fly, honey bee and mosquito sequence databases. A Drosophila paralog of the 5-HT_2βreceptor, 5-HT_2αDro [originally 5-HT_2Dro (see Tierney, 2001; Clark et al., 2004)] binds strongly to α-Me-5-HT, 2-Me-5-HT, ritanserin and methysergide(Colas et al., 1995), all of which we found to be functionally active at 5-HT_2βPan and 5-HT_2βPro. These drugs may therefore, be 5-HT₂-specific antagonists in arthropods. N-acetyl-5-HT and ketanserin were completely inactive at crustacean 5-HT_2β but have very high binding constants at 5-HT_2αDro. Effective binding of a drug to a receptor may not directly reflect that drug's functional properties at the receptor. 5-HT_2β and 5-HT_2α may therefore be more or less similar, functionally,than suggested by these comparisons. Looking beyond arthropods, several 5-HT₁ and 5-HT₂ type receptors from mollusks and nematodes, also bind methiothepin or ritanserin(Angers et al., 1998; Gerhardt et al., 1996; Hamdan et al., 1999; Olde and McCombie, 1997; Sugamori et al., 1993). Similarly, (+)butaclamol binds a variety of 5-HT receptor subtypes from diverse species (Hamdan et al.,1999; Olde and McCombie,1997; Saudou et al.,1992). In sum, the antagonist profile of 5-HT receptors appears to be well conserved among crustaceans but this may not extend between phyla.

Crustacean 5-HT_1α receptors are highly resistant to antagonists. Even the potent and selective mammalian 5-HT_1A blocker WAY100635 (Hoyer et al., 2002)was ineffective at 5-HT_1αPan. Prazosin, which functionally blocks 5-HT_1Dro (Saudou et al.,1992), also had no effect on crustacean 5-HT_1α. Many of these putative antagonists block functional activation of mammalian(Hoyer et al., 2002) and invertebrate (Barbas et al.,2002; Hobson et al.,2003; Li et al.,1995; Saudou et al.,1992; Witz et al.,1990) 5-HT receptors. Although some of these may act as allosteric modulators and therefore be sensitive to the high variability of the N-terminal tail amongst 5-HT receptors(May et al., 2004), several of the drugs do efficiently displace radioligands at arthropod receptors(Tierney, 2001) and therefore presumably bind at or near the ligand binding pocket. For none of these to be capable of blocking 5-HT activation of crustacean 5-HT_1α is unexpected. 5-HT_1α may have an extraordinarily high affinity and selectivity for 5-HT such that the antagonists are not able to overcome the binding of and/or conformational changes elicited by 5-HT. Radioligand studies done on other arthropod 5-HT_1α receptors, however, do not fully support this hypothesis as some putative antagonist drugs bind more strongly than 5-HT (Colas et al.,1995; Saudou et al.,1992). Alternatively, the exogenously expressed crustacean 5-HT_1α may be in a hyperactive state that overcomes any antagonist effects. This might explain the high response to almost all the putative agonists we tested.

We previously showed that 5-HT_1α was extensively expressed in thoracic ganglia of crayfish and Macrobrachium rosenbergii, the giant freshwater prawn, in similar patterns(Sosa et al., 2004; Spitzer et al., 2005). Crustacean 5-HT receptors may therefore be conserved in their expression patterns as well as in their molecular structure and function. In crayfish,5-HT_1α is observed in somata and the neuropil throughout the central nerve cord. It is also localized to processes surrounding the nerve roots, to superficial flexor muscles of the abdomen and to processes on the vasculature (Spitzer et al.,2005). The 5-HT_1α receptor therefore provides numerous targets for 5-HT modulation in the crayfish.

Several of the drugs characterized here have been used to investigate physiological mechanisms of 5-HT neuromodulation in crustaceans. Based on the pharmacological characterization of crustacean 5-HT_2β and 5-HT_1α we can now hypothesize which receptors might be involved in mediating specific effects and behaviors. In the lateral giant neuron of crayfish, mCPP elicited the inhibitory component of the 5-HT response (Yeh et al., 1997)indicating that 5-HT_1αPro may be responsible. The facilitatory component of the 5-HT response was elicited by α-Me-5-HT(Yeh et al., 1997), an effective agonist of both 5-HT_2βPro and 5-HT_1αPro. One or both of these receptors could therefore contribute to facilitation of the circuit. In freely behaving crayfish, mCPP injection mimicked the effects of 5-HT on posture(Tierney et al., 2004; Tierney and Mangiamele, 2001)indicating that 5-HT_1αPro may mediate 5-HT effects on posture. Crayfish abdominal muscles, essential to posture, are modulated by 5-HT at the motor neurons and the neuromuscular junction(Strawn et al., 2000). 5-HT_1αPro receptors are localized on abdominal superficial flexor motorneurons (Spitzer et al.,2005) and could mediate this effect. In addition, 5-CT enhances agonistic behavior in Procambarus(Tierney and Mangiamele,2001), indicating that 5-HT_2βPro and/or 5-HT_1αPro could mediate 5-HT signals eliciting these behaviors. Because posture is an essential component of agonistic behavior,these receptors may therefore contribute to different aspects of the agonistic behavioral program (Edwards and Kravitz,1997; Huber et al.,1997; Kravitz,2000). Although 5-HT_1αPro is expressed by neurons throughout the crayfish eyestalk (Spitzer et al., 2005), it may not mediate 5-HT-elicited hyperglycemia as this was not mimicked by mCPP (Lee et al.,2000). 5-HT_2βPro may, however, be involved as the response was evoked by both 5-CT and 8-OH-DPAT. In the stomatogastric nervous system of the crab, Cancer borealis, cinanserin blocked 5-HT-mediated acceleration of the pyloric rhythm (Zhang and Harris-Warrick, 1994) suggesting that the crab 5-HT_2β ortholog may play a role in modulation of pyloric frequency. Indeed, we have identified roles for 5-HT_2βPan,5-HT_1αPan and an unidentified 5-HT receptor in modulation of the pyloric rhythm in Panulirus using the drugs identified here (N.S. and D.J.B., unpublished).

At least three crustacean 5-HT receptors (5-HT_1β,5-HT_2α, 5-HT₇) remain uncharacterized(Clark et al., 2004; Tierney, 2001) and, as we demonstrate in comparing crustacean 5-HT_2β and 5-HT_1α receptor pharmacology, many drugs are active at multiple 5-HT receptor subtypes. Uncharacterized 5-HT receptors could therefore be contributing to observed effects. Especially when using a single agonist drug to elicit a `5-HT effect', the possibility of activating multiple 5-HT or other aminergic receptors must be considered. In addition, radioligand displacement assays of Drosophila 5-HT_1α and 5-HT_1β, indicated that arthropod 5-HT receptor paralogs can have similar pharmacological properties(Saudou et al., 1992); their effects may therefore be difficult to tease apart even using several drugs. Combinations of drugs were, however, used to assign specific modulatory roles to 5-HT_2β, 5-HT_1α and an uncharacterized receptor in the stomatogastric nervous system of Panulirus (N.S. and D.J.B., unpublished). Profiling of additional receptor subtypes, the use of multiple characterized drugs and expression studies should allow greater confidence in assigning specific receptor types to behavioral effects.

In conclusion, we have shown that crustacean 5-HT_1αreceptors are functionally similar to homologous receptors across multiple vertebrate and invertebrate species; furthermore, 5-HT receptors are conserved in protein structure and signaling properties. For 15 drugs that we were able to test at both receptors in both species, the effects on Panulirusand Procambarus receptors were very comparable. Furthermore, the effects of 19 additional drugs that could only be compared for one receptor type were also similar between the two species. These comparisons suggest that the pharmacological profiles of 5-HT receptor orthologs between these two infraorders are relatively well conserved. Finally, we have identified drugs specific to 5-HT receptor subtypes and thus developed a pharmacological toolset for use in crustacean physiological preparations.

We are grateful to Merry Clark for training and insightful discussions and thank Jin Kim and Ping Jiang for excellent technical support. N.S. was a scholar in the Brains and Behavior Program and the Center for Behavioral Neuroscience at Georgia State University. This work was funded by NIH NS38770 to D.J.B., NIH MH62167 to D.H.E. and NSERC PGS-D2-304455-2004 to N.S. Partial support came from the NSF under agreement IBN-9876754 to the Center for Behavioral Neuroscience.